Incident management · Production

How Datadog Built Bits AI SRE: An Autonomous Incident Investigation Agent That Reduces Time to Resolution by Up to 95%

The problem

As distributed systems grow more dynamic and complex, production incidents span more services, involve noisier signals, and generate larger volumes of telemetry data, making it hard for on-call engineers to find root causes quickly.

First attempt

Early SRE agents performed many tool calls and summarized all telemetry at once, causing token counts to scale linearly with complexity, which degraded model performance and led to incorrect root cause identification when noisy signals distracted the summarization prompt.

Workflow diagram · grounded in source
1
Monitor alert triggers investigation
trigger
“automatically investigates incidents and monitor alerts by autonomously reasoning over complex telemetry data”
2
Hypothesis formulation
ai_action
“Formulate hypotheses about the root cause”
3
Targeted query validation
validation
“Validate or reject hypotheses using data from targeted queries”
4
Recursive sub-hypothesis exploration
ai_action
“This version of the agent recursively generates deeper root cause hypotheses until it exhausts the search space, allowing for deeper, more insightful investigations into an alert”
5
Audit-ready root cause analysis
output
“producing audit-ready root cause analyses in minutes”
Reported outcome

Bits AI SRE decreases time to resolution by up to 95% and has received overwhelmingly positive feedback from customers who observed reduced time to root cause detection for complex incidents.

Reported metrics
Time to resolutionup to 95%
Time to root cause detectionreduced time to root cause detection for complex incidents
Agent capabilities improvementsignificantly improved over the past year
Reported stack
Bits AI SRE
Source
https://www.datadoghq.com/blog/building-bits-ai-sre/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Bits AI SRE decreases time to resolution by up to 95% and has received overwhelmingly positive feedback from customers who observed reduced time to root cause detection for complex incidents.

What tools did this team use?

Bits AI SRE.

What results were reported?

Time to resolution: up to 95%; Time to root cause detection: reduced time to root cause detection for complex incidents; Agent capabilities improvement: significantly improved over the past year (source-reported, not independently verified).

What failed first in this deployment?

Early SRE agents performed many tool calls and summarized all telemetry at once, causing token counts to scale linearly with complexity, which degraded model performance and led to incorrect root cause identification…

How is this incident management AI workflow structured?

Monitor alert triggers investigation → Hypothesis formulation → Targeted query validation → Recursive sub-hypothesis exploration → Audit-ready root cause analysis.